Final Report Management and restoration analysis for Kankakee sands black oak sand savanna

Grant Agreement IDNR ‐08‐031W

John W. Groninger Dept. Forestry‐4411 Southern Illinois Univ. Carbondale, IL 62901 Phone 818‐453‐7462

Timeframe Start 7/16/07‐End 6/30/08

Final Report prepared by: John W. Groninger Dept. Forestry‐4411 Southern Illinois Univ. Carbondale, IL 62901 Phone 818‐453‐7462

Project objectives (from proposal):

Kankakee Sands encompasses the largest concentration of high quality black oak sand savannas in the Midwest. This area provides a critical habitat for many unique plants and animals. For example the Red‐ headed woodpecker (Melanerpes erythrocephalus), the only known population of Yellow wild indigo (Baptisia tinctoria), Regal fritillary (Speyeria idalia), eastern most population of Plains pocket gopher (Geomys bursarius), and Primrose violet (Viola primulifolia) are only a few of the many species dependent on the black oak sand savannas. According to scientists of The Nature Conservancy (TNC), an additional 8,000 acres is needed to support the black oak sand savanna system. The large size and concentration of high quality black oak sand savannas of Kankakee Sands provides managers a valuable opportunity to develop management guidelines for these systems.

The herbaceous component of savanna systems has been widely studied. However, the tree component, dominated by the relatively short‐lived black oak, has received little attention. Presently, many of these areas are overstocked and otherwise in poor condition. Even among high quality systems, long‐term stability is inferred, but not known.

If funded, this project will provide specific recommendations for the long‐term management of the tree component of black oak sand savannas. Information needed to develop these guidelines will be based on measurements of structure, dynamics, and vegetation characteristic of high quality and degraded black oak sand savannas. Considerable attention will be paid to the actual age, structural stability and likelihood of survival of; a. presently dominant black oak individuals, b. small trees in a position to assume dominance following the demise of these individuals, and c. the status of regeneration. Site‐ specific recommendations for fire regime maintenance, regeneration needs and density management will be developed for both high quality and degraded savannas. The results of this project will also assist resource managers improve the prospects for successful black oak sand savanna stewardship throughout Illinois. Completed project description (complete description attached)

Summary Management and restoration of black oak dominated sand savannas rely on historic vegetative descriptions (settler accounts, surveyor notes, aerial photographs). It is commonly assumed that fire alone maintains savanna structure and composition, however little is known about the specific fire frequency needed to maintain these systems. The objective of this study was to quantify and correlate characteristics of stand structure with fire history of the Kankakee Sands savannas in northeastern Illinois. Fire history chronologies were determined through dendrochronological methods from 287 dated fire scars identified on 60 black oak (Quercus velutina) trees located throughout four study sites. Stand structure was characterized in 30 circular plots (0.04 hectares) per study site that were placed at 25‐meter intervals along randomly established permanent line transects in the summer of 2007. Stand structure analysis consisted of the development of tree age‐size relationships among presently dominant and suppressed trees in relation to fire history. Variations in stand structure were strongly related to fire dynamics among the four study sites. Specifically, components of stand structure such as tree density (n = 118; r = 0.440; p < 0.0001), basal area (n = 118; r = 0.347; p < 0.0001), and total woody stem density (n = 118; r = 0.417; p < 0.0001) all increased as a function of fire free interval. The population of black oak, the dominant canopy tree (I.V. 300 = 244.66) appeared to be structurally unstable based on the historic fire regime of the four sites. Fire intervals less than two years maintained conditions of openness, as was referenced to 1939 historic aerial photographs, but eliminated potential future canopy trees. Under these conditions, we predict a dramatic shift in community structure toward prairie vegetation, as no smaller trees were present to assume canopy dominance. Fire intervals greater than two years were associated with transition to closed canopy forests. Other management considerations, including grazing and or selective cutting, are proposed to balance historic canopy openness and promote regeneration of characteristic savanna species. Images

Kankakee sands black oak savanna

Kankakee sands black oak savanna

Black oak cross‐section with fire scars

Collecting black oak cross sections Species benefitting from this project include…

Red‐headed woodpecker (Melanerpes erythrocephalus) Yellow wild indigo (Baptisia tinctoria) Regal fritillary (Speyeria idalia) Plains pocket gopher (Geomys bursarius) Primrose violet (Viola primulifolia) Schizachyrium scoparium (little blue stem) Andropogon gerardi (big blue stem) Heterotheca spartea (porcupine grass) Carex pensylvanica (pen sedge) Koeleria macrantha (June grass) Amorpha canescens (lead plant) Aster linariifolius (stiff‐leaved aster) Comandra umbellata (bastard toadflax) Liatris aspera (blazing star), Corylus americana (American hazel) Rhus sp. (sumac) Salix humilis (prairie willow) Rosa caroliniana (Carolina rose)

Specific audience affected

Visitors to Kankakee sands Managers and users of black oak sand savannas elsewhere in Illinois

Products

‐MS thesis of Cody Considine (in progress)

‐Presentation at Ecological Society of America meeting: Considine, C.D., J.W. Groninger, C.M. Ruffner, S.G. Baer, and M.D. Therrell. 2008. Integrating fire history and stand structure: Implications for black oak sand savanna management and restoration.Ecological Society of America Annual Meeting, Auguest 8, Milwaukee, WI.

‐SIUC press release:

Hello from Southern Illinois University Carbondale. We are sending you (1) news release. Today’s headline is:

1. Research may help save Kankakee Sands savanna

Research may help save Kankakee Sands savanna By K.C. Jaehnig

CARBONDALE, Ill. ‐‐ Flame plus cow usually equals nothing more than a tasty burger, but when it comes to saving one of the nation’s few remaining savannas, that combination might prove more than a value meal. Research at the Kankakee Sands, a Nature Conservancy preserve straddling the Illinois‐ Indiana state line near Kankakee, suggests that burning and grazing could work together in maintaining black oak savanna ecosystems, which require a tricky balance of trees and grasslands to be at their best. Funding for the work came from The Nature Conservancy, the U.S. Department of Agriculture and the Illinois Department of Natural Resources Wildlife Preservation Fund. “Because of their open conditions, these areas are very rich in flowers and other vegetation, birds and animals, “ said John W. Groninger, who teaches in Southern Illinois University Carbondale’s forestry department and supervised the research. “Without fire, they become overgrown with trees, which provide too much shade for other vegetation to grow. But we found that the areas that burned enough to maintain a somewhat open structure still had undergrowth coming in. It looks like grazing is the missing element.” Covering roughly a third of the Kankakee Valley, the Kankakee Sands region is a legacy of the last ice age when floodwaters from prehistoric lakes left behind gravel and sand as they roared through. Swept by the wind, the sand formed dunes, which offered some protection to fast‐sprouting, drought‐tolerant black oaks springing up in the protected swales. Over time, this unusal blend of oak and grass became home to, among other species, redheaded woodpeckers, regal fritillary butterflies, plains pocket gophers, yellow wild indigo and orange fringed orchids, all of which nowadays struggle to survive. “Savannas are like prairies with scattered trees,” said Cody D. Considine, a restoration ecologist with The Nature Conservancy, who is basing his master’s thesis on his research in this area. “Because some species require savanna habitat, this area is critical for many unique plants and animals. Redheaded woodpeckers, for example, can’t live in closed canopy forests or open prairies. They need a mix of grassland and trees.” Over the years, development and row crops have destroyed the once‐plentiful savannas, reducing millions of acres throughout the Midwest to little more than a few thousand. The Kankakee Sands area contains the largest concentration of black oak savannas not just in the Midwest but in the country. Maintaining what’s there and adding to it plays a crucial role in protecting the plants and wildlife that depend on this type of habitat. Fire has always served as a key player in keeping these savannas relatively open. But too much fire harms more than it helps. “The larger trees can survive, but younger trees are often killed,” Considine said. “Even if they can resprout, with frequent fires the saplings are unlikely to grow into trees.” Considine decided to track the fire history of four sites in the Kankakee Sands by sectioning some tree trunks and analyzing their fire scars. That information would suggest how often managers could perform prescribed burns to maintain openness without killing the younger generation. His idea worked — up to a point. “The areas that were turning into forest were the ones that had been burned the least, and the more open areas were the ones that had been burned the most,” Considine noted. But there was a problem. “Fire intervals of less than two years maintain the open structure but absolutely no regeneration takes place,” Considine said. “When all the larger trees die, you’d have a shrub prairie, not a savanna. “But with intervals over three years, the savanna turns into forest. That’s where grazing comes in. We’re speculating that you wouldn’t have to burn as much to maintain that open structure while still regenerating trees.” Grazing removes grass and other fuel that makes fire burn hotter, upping survival odds for the younger trees. Bison used to perform that task; cows could do it now. “This was somewhat surprising to us, because with our background in forestry, we’re used to cattle being a destructive force as far as mature trees go,” Groninger said. “But we’re learning that under certain conditions, they can play an important role in the regeneration process. If we can find ways of producing cattle under conditions that sustain natural areas, we could maintain both rural economies and preservation activities.” Further research awaits funding and students with an interest in performing it. But Considine, who plans to tie up the loose ends on his thesis in September, isn’t waiting. He and his fiancée have bought a tract of land near Dixon. They’ll be burning and running cows and learning how it all works out.

Project expenditures‐ Already submitted.

Funding from other sources SIUC match $3474 The Nature Conservancy $4000

1

---DRAFT: DO NOT CITE---

INTEGRATING FIRE HISTORY AND STAND STRUCTURE:

IMPLICATIONS FOR BLACK OAK DOMINATED SAND SAVANNAS

------DRAFT: DO NOT CITE-----

Summary

Management and restoration of black oak dominated sand savannas rely on historic vegetative descriptions (settler accounts, surveyor notes, aerial photographs). It is commonly assumed that fire alone maintains savanna structure and composition, however little is known about the specific fire frequency needed to maintain these systems. The objective of this study was to quantify and correlate characteristics of stand structure with fire history of the Kankakee Sands savannas in northeastern

Illinois. Fire history chronologies were determined through dendrochronological methods from 287 dated fire scars identified on 60 black oak (Quercus velutina) trees located throughout four study sites. Stand structure was characterized in 30 circular plots (0.04 hectares) per study site that were placed at 25-meter intervals along randomly established permanent line transects in the summer of 2007. Stand structure analysis consisted of the development of tree age-size relationships among presently dominant and suppressed trees in relation to fire history. Variations in stand structure 2

were strongly related to fire dynamics among the four study sites. Specifically,

components of stand structure such as tree density (n = 118; r = 0.440; p < 0.0001), basal area (n = 118; r = 0.347; p < 0.0001), and total woody stem density (n = 118; r =

0.417; p < 0.0001) all increased as a function of fire free interval. The population of black oak, the dominant canopy tree (I.V. 300 = 244.66) appeared to be structurally

unstable based on the historic fire regime of the four sites. Fire intervals less than two years maintained conditions of openness, as was referenced to 1939 historic aerial photographs, but eliminated potential future canopy trees. Under these conditions, we

predict a dramatic shift in community structure toward prairie vegetation, as no smaller

trees were present to assume canopy dominance. Fire intervals greater than two years

were associated with transition to closed canopy forests. Other management

considerations, including grazing and or selective cutting, are proposed to balance

historic canopy openness and promote regeneration of characteristic savanna species.

Introduction

Importance of restoration and management of black oak sand savannas

Currently, oak savanna management and restoration efforts rely on historic vegetative

descriptions (Gleason 1913, Anderson & Fralish 1975, Cutter & Guyette 1994), coupled

with few experimental studies to create assumptions of ecologic relationships (Bowles

and Hutchinson 1994). Researchers have suggested that fire alone maintains stand

structure and composition (Tester 1989, Apfelbaum and Haney 1991, Packard 1991

and 1993, Peterson and Reich 2001). Of the remaining savannas of the Midwest, the

majority of them, credited to fire suppression have transitioned into closed canopy

forests (Auclair 1976), thus require other aggressive approaches, such as mechanical 3 tree removal to restore stand structure and composition (Vogl 1964, White 1983,

Peterson and Reich 2001, Nielson et al. 2003). Most resource managers and scientists agree that these efforts (i.e. fire and thinning) are essential to improve these ecosystems; yet, there is a lack of understanding on fire frequency effects pertaining to stand structure, dynamics (Peterson and Reich 2001), and herbaceous composition

(Leach and Givnish 1999).

Oak savannas have been a significant vegetation community throughout the

Midwest landscape for the last 8,000 years (McClain 1991). Landscape heterogeneity such as topographic barriers (rivers, lakes, and bluffs) served as protection from landscape prairie fires, therefore allowing savannas to exist within the tallgrass prairie region (Grimm 1984, Whitney and Steiger 1985, Leitner et al. 1991). Midwest oak savannas are estimated to have covered between 11-13 million hectares at the time of

Euro-American settlement; encompassing Illinois, Iowa, Minnesota, Wisconsin,

Missouri, Michigan, Indiana, and Ohio (Nuzzo 1986). Currently, less than 0.02 % of

Midwest oak savanna remains, this consists mostly of dry to dry-mesic savannas with no records of any intact deep soil savanna as of 1985 (Nuzzo 1986). As with nearly all

Midwest natural communities, the primary loss of oak savanna is due to the conversion of agriculture and development (Nuzzo 1986).

Not only is the basis of oak savanna management and restoration solely a result of their conservation importance in terms of species habitat, but just as important are their ecosystem services. According to Costanza et al. (1997), oak savanna ecosystem services are valued between $232 ha-1 yr-1 to $302 ha-1 yr-1 or approximately $906 x 109 in total (1996 U.S. dollars). These services include gas regulation, climate regulation, erosion control, soil formation, waste treatment, food production, and recreation

(Costanza et al. 1997). In combination of the economic and conservation importance, 4

many resource agencies throughout the Midwest have focused on improving oak

savanna habitat. Even with increased awareness and extensive restorations within the

past few decades, few studies have investigated how much fire is necessary to sustain

savanna stand structure (Peterson and Reich 2001) while also retaining and enhancing

herbaceous richness well into the future.

Eastern Sand Savanna Ecology

Of the dry savannas remaining in the Midwest, eastern sand savannas (Haney and

Apfelbaum 1995) are the most common (Will-Wolf & Stearns 1999). Eastern sand

savannas are classified as dry soil savannas located throughout the upper Midwest in

northern and central Illinois, southern Wisconsin, southern Michigan, northwestern

Indiana, northwestern Ohio, eastern Iowa, and southern Minnesota (Gordon 1966,

Homoya 1994, Coles & Taylor 1995, Johnson & Ebinger 1995, Haney & Apfelbaum

1995) (Figure 1).

Eastern sand savannas occur on formations of sandy, low nutrient, dry soils of glacial moraines, sandy lake beds, outwash plains, and dune systems (Nuzzo 1986,

Will-Wolf &Stearns 1999). Soils are further characterized as being excessively drained

with low cation exchange capacities and less than 2% organic matter (Haney &

Apfelbaum 1995). As with all types of Midwest oak savannas, the fire tolerant Quercus

species thrive. Quercus velutina (black oak) is the dominant tree of eastern sand

savannas. While other oaks including Quercus alba (white oak), Quercus ellipsoidalis

(northern pin oak), and Quercus macrocarpa (bur oak) may be present along with other non oak species including Prunus serotina (black cherry), Sassafras albidum

(sassafras) and Robinia pseudoacacia (black locust) depending on environmental gradients and disturbance history within a site (Whitford & Whitford 1971, Johnson & 5

Ebinger 1992, Ebinger et al. 2006).

Savannas are considered transitional areas (Anderson 1998) or ecotones

(Mendelson et al. 1992), which are habitats that are not only a physiognomic fusion of prairie and forest, but share similar species of each habitat as well. Therefore, the mosaic of scattered, open grown, fire tolerant trees create heterogeneous microsite conditions allowing for rich species diversity and substantial turnover in species composition of forbs and grasses (Anderson 1998, Leach and Givnish 1999). Although overstory species of eastern sand savannas are not very diverse, ground layer vegetation associated within these systems are extremely diverse (Phillippe et al. in review) with species especially adapted to xeric conditions (Ebinger et al. 2006).

Common graminoid, forb, and shrub species associated in eastern sand savannas systems include Schizachyrium scoparium (little blue stem), Andropogon gerardi (big blue stem), Heterotheca spartea (porcupine grass), Carex pensylvanica (pen sedge), and Koeleria macrantha (June grass) Amorpha canescens (lead plant), Aster linariifolius

(stiff-leaved aster), Comandra umbellata (bastard toadflax), Liatris aspera (blazing star),

Corylus americana (American hazel), Rhus sp. (blackberry), Salix humilis (prairie willow) and Rosa caroliniana (Carolina rose) (Johnson & Ebinger 1992, 1995, Phillippe et al.

2003). However, distributions of herbaceous species diversity found throughout these systems can be extremely variable depending on past land use history and tree density

(Phillippe et al. in review).

Disturbance ecology

Disturbances are primarily responsible for creating and maintaining the unique structure of eastern sand savannas because otherwise in a temperate climate where there is sufficient precipitation for tree growth, these systems transition into closed canopy 6

forests (Faber-Langendoen & Davis 1995, Carter & Johnson 1986). Disturbances

credited as the “keystone” ecologic processes within these systems are fire, grazing

(Anderson and Bowles 1999), and human impacts particularly Native American’s

frequent use of fire (Guyette et al. 2006) and tree cutting. Fire has been well

documented throughout the past century as integral force in influencing the structure

and dynamics of eastern sand savannas (Gleason 1913, Curtis 1959, McPherson

1997).

Methods

Study Site

This study was conducted within the wooded areas of Pembroke Township (41° 04’ N,

87° 37’ W) located in the vast moraine and sand deposits of Kankakee County, Illinois.

Elevations at the site range between 203 meters to 212 meters above sea level. The

thirty year (1971-2001) mean annual temperature is 9.9°C, averaging between highs of

23.6°C (July) to lows of -5°C (January). The thirty year (1971-2001) mean annual precipitation is 98.0 cm, May having the greatest amount at 11.5 cm (Midwest Regional

Climate Center 2008). Pembroke Township is located at the western edge of the former glacial lake, Lake Watseka, which was drained during the Kankakee Torrent about

14,500 years ago (Phillippe et al. 2008). During the last glacial retreat, massive amounts of glacial melt-water cut “channels” through the existing glacial deposits that acted as dams resulting in massive flooding in the Kankakee, Vermilion, and Fox River valleys. This event scoured the Illinois River basin to bedrock, transporting sand and gravel along the Kankakee and Illinois River valleys. Overtime, these exposed sand deposits were transported by the prevailing westerly winds which created the series of dunes and swales found throughout Pembroke Township (Dawson 2007). All the study 7

plots were established on Oakville soils that are characterized as excessively drained,

have a dark grayish brown surface horizon, with 152.4 cm of fine sand extending into

the soil profile (Paschke 1979).

The extensive sandy soils have deemed this area unfavorable for row crop

agriculture. Although, during the mid 1800’s to early 1900’s grazing cattle was very

predominant through out the area, supplying the growing city of Chicago, located just 65

miles to the north (Dawson et al. 2007). Other post settlement, historic land use

disturbances include frequent fires, grazing, and sporadic tree cutting for firewood have

been speculated to be the main components in maintaining the quality of these

savannas (Phillippe et al. 2008).

Coupled with the unique sand dune habitat and consistent disturbance

throughout pre and post settlement, this region supports very distinct and unique animal

and plant populations. Approximately 22 rare plant species can be found in the

Pembroke savannas including the largest Illinois population of Platanthera ciliaris

(orange-fringed orchid), the only Illinois population of Baptisia tinctoria (yellow false

indigo), Viola primulifolia (primrose violet) to name a few. Distinct reptilian, and

mammalian species can also be found in Pembroke such as Terrapene ornatata (ornate

box turtle), Ophisaurus attenuatus (western glass lizard), Cnemidophorus sexlineatus

(Six-lined racerunner), and the only Illinois gopher species Geomys bursarius (plains pocket gopher). Populations of Midwest savanna’s keystone species, Melanerpes erythrocephalus (Red-headed Woodpecker) (Brawn 1998, Blood 2002) are thriving at

Pembroke (Brawn ?), even though populations are declining throughout their range at an annual rate of 2.5% (Sauer et al. 2001). A recent study by Phillippe et al. (2008) inventoried the woody overstory and ground layer vegetation of Pembroke Township in

2002. They concluded the overstory was not very diverse, with Quercus velutina (black 8

oak) being the dominant tree species with few occurrences of Quercus alba (white oak) and even fewer occurrences of Quercus palustris (pin oak). However, ground layer vegetation was very diverse, totaling 574 species. Also, based on their analyses of the

1939, 1968, and 1988 aerial photographs, some areas have transitioned into open woodlands and closed canopy forests. The entire region is comprised of 10,000 acres of black oak sand savanna located in Illinois and Indiana.

The savannas located in the Pembroke Township are recognized by numerous conservation agencies as very high quality sand savannas. In 1978 the Illinois Natural

Areas Inventory identified these savannas as state significant natural areas (White

1978). The Nature Conservancy (200?) considers these savannas as an important ecoregion site and the United States Fish and Wildlife Service (1999) have labeled these savannas to be the greatest concentration of high-quality black oak sand savannas in the nation. However, not all these areas are protected and some areas are more degraded than others. The Illinois and Indiana chapters of The Nature

Conservancy and Department of Natural Resources are the primary land managing agencies protecting approximately 6000 acres of black oak sand savanna. Management efforts have primarily been land acquisition and prescribed burning with some thinning of undesirable trees species (Robinia pseudoacacia), exotic weed control, and fence construction to keep all terrain vehicles out.

Experimental Design

Stand Structure

Permanent study plots were established on five sites within Pembroke Township by the

Illinois Natural History Survey (INHS) in the summer of 2002. Four of those sites were re-inventoried for this study in the summer of 2007. INHS inventoried and recorded very 9

detailed lists of ground layer vegetation of each plot located among the sites. Therefore,

references between herbaceous vegetation with stand structure and fire history can be

analyzed; for this reason was the premise for re-inventorying their plots. Within the

sites, thirty circular (0.04 ha) plots were placed at twenty-five meter intervals along

permanent line transects randomly established along cardinal compass directions

(Phillippe et al. 2003). The locations of every plot were recorded by taking GPS

coordinates.

Stand structure characteristics were determined through the identification and measurements (dbh tape) of all living and dead standing trees ≥ 10.0 cm diameter at

breast height (dbh) and all living oak grubs and saplings < 10.0 cm dbh in each plot.

Percent cover of all woody species within each plot was estimated as well. Tree health

measurements were additionally taken characterize stand health of these sites and to

provide baseline information for future studies. These measurements include tree status

(dead, alive, or leaning), total live crown height (m), lower live crown height (m), crown

width (N&S, E&W) (m), number of multiple stems, percent canopy loss, number of dead

branches ≥ 7.6 cm, visible fire scars present (yes or no), hollow (yes or no), root

suckers present (yes or no), visible decay on bole of tree present (yes or no).

This information was used to calculate both the plot and site level of each

species and combined total of live stem density of trees and saplings (stems/ha), tree

basal area (m2/ha), importance value (IV), relative dominance, relative density,

averages computed for diameter (cm), crown width (m), crown height (m), number of

dead branches ≥ 7.6 cm, percent canopy loss, crown volume (m3), volume of healthy

crown(m3), percentage of trees hollow, and average size of visible wounds.

To quantify site age structure, increment cores were extracted from 154 trees

located in the four sites. The tree cores were taken back to Southern Illinois University 10

Carbondale for dendrochronological analysis. The tree cores were glued on core mounts where they were sanded using progressively finer sand paper (150-600 grit).

Ages were determined under a stereomicroscope (10x) after the dendrochronological analysis was performed on the cross-sections for the fire history analysis portion. Ages could only be confirmed on 63 of the 154 cores due to extensive decay. Age distribution graphs were created from the combined tree cores and cross-sections ages.

Fire History

In January of 2008, sixty Quercus velutina (black oak) cross-sections, fifteen from each site were collected from all four sites inventoried in the summer of 2007

(Bentley/Crawford, Big Dune, Leesville, and Mskoda). Living trees were selectively chosen by first visually inspecting the tree, while some trees were cored with an increment corer or sounded with an axe to ensure the tree was not hollow. A chainsaw was used to initially cut each tree down, then cross-sections were removed, however two cross-sections were collected from stumps of trees that were illegally removed from the Leesville site. Geographic locations (GPS coordinates) and orientation of visible fire scar were recorded for each tree used in the analysis.

Next, all the cross-sections were taken back to the tree ring laboratory at

Southern Illinois University Carbondale for dendrochronological analysis. In attempts to safeguard against any foreign insects or diseases, samples were left outside for two days where temperatures were below freezing. Prior to sanding, the samples were then placed in the green house to dry for two-weeks. After further visual examination in the lab, five cross-sections were omitted from the analysis due to extensive decay. Surfaces of each sample were planed with an electric planer, sanded using progressively finer sandpaper (60 – 1200 grit), and then completed with fine steel wool to expose cell 11 structure of the annual rings under a 10x stereomicroscope. Once the samples were prepped, visual cross-dating procedures called skeleton plots were created by identifying signature years through graphically expressing the width of each annual ring of every sample (Stokes & Smiley 1968). Next, the annual rings of thirty-five sections were measured with a stereomicroscope using the computer program Measure J2X

(Voorhees 2000) and a Velmex measuring system, the quality-control program

COFECHA (Holmes 1983, Grissino-Mayer & Holmes 1993) was used to detect false or missing rings, and a master chronology was created from the computer program

ARSTAN (Cook and Holmes 1984). The samples and dating chronology was then cross-dated with a Quercus alba chronology from Kankakee State Park within 33 km of all study sites (Duvick 1980, World Data Center for Paleoclimatology 2008).

Once the annual rings of each sample were dated correctly, fire scars were identified and dated. Fire scars were identified by the presence of charcoal, vascular cambium injury, and or a disruption of an annual ring that showed healing in later years

(Wolf 2004). Calendar dates of fire scars were assigned to the seasonality and year of cambial injury with dormant season fires dated to the following growing season (Baisan

& Swetnam 1990, Guyette et al. 2006). Seasonality of the scar is based on the injury location within the annual ring and designated as early, middle, late growing, or dormant season wood. Finally, the fire scar dates and seasonality information were inputted into the fire history program, FHX2 (Grissino-Mayer 2001) to compute statistical analyses on fire occurrences and create tree and composite fire scar chronology graphs. Weibull median fire intervals (WFI) and mean fire intervals (MFI) were calculated for all four sites. The analysis approach to determine MFI and WFI was set at 1930 when at least one tree of each site was present. The superposed epoch analysis (SEA) was used to 12 determine the relationship between the frequencies of fire events with the mean Palmer

Drought Severity Index for the period 1930 to 2007.

Kankakee County Cattle Population

The United States Department of Agriculture National Statistics Service (2008) was used to determine the population of non-feedlot cattle from 1926 to 2007 in Kankakee

County Illinois.

Statistical Analyses

Differences between all four sites stand structural characteristics at both the plot level and means of each site were analyzed by the Least Square Means function in a 1-way

ANOVA using the glm procedure in SAS (SAS Inc. 2003). To determine if relationships exists between fire history and current stand structure, Pearson’s correlation coefficients were used to test for correlations between response variables in SAS (SAS 2003).

Results

Stand Structure

Results of the stand structure inventory of the four sites combined indicates that the dominant tree species, Quercus velutina (black oak) had the greatest frequency (80%), density (111 trees/ha), basal area (7 m2/ha), resulting in a high importance value

(244.60 300) (Table ?). The other two species present throughout the sites included

Quercus alba (white oak) and Quercus palustris (pin oak). White oak was not nearly as

common as black oak, found infrequently in three of the four sites (Bentley, Leesville, and Mskoda) with less density (26 trees/ha), basal area (1.19 m2/ha), and frequency

(18%), and thereby a low importance value (51.28). Pin oak comprised only a few plots 13 in Leesville, accounting for only 2% of the frequency, 2.29 trees/ha, 0.05 m2/ha of basal

area.

Stand structural characteristics varied greatly between the four study sites. Big

Dune had the lowest stand density (42.5 trees/ha) while Bentley/Crawford had the

greatest density (247.50 trees/ha) with Leesville (107.50 trees/ha) and Mskoda (158.33

trees/ha) comprising similar tree densities (F = 26.39; df = 3, 110; P < 0.0001) (Table ?).

Bentley/Crawford also accounted for the greatest tree basal area (13.65 m2/ha) while

Big Dune (4.93 m2/ha), Leesville (7.05 m2/ha), and Mskoda (7.4 m2/ha) had similar

basal areas (F = 17.93; df = 3, 110; P < 0.0001) (Table ?). Oak saplings/grub density and total woody sapling and shrub percent cover was also relatively different among the four sites. Big Dune (559 stems/ha) had the lowest oak sapling/grub density while

Leesville (1827 stems/ha) had the greatest density with Mskoda (1079 stems/ha) and

Bentley/Crawford (1060 stems/ha) containing similar oak sapling/grub densities (F =

5.14; df = 3, 110; P < 0.0023) (Table ?). The combined total of woody vegetation, Big

Dune (599.13 stems/ha) comprised the lowest mean density while Leesville (1975.40

stems/ha) had the highest with Mskoda (1233.03 stems/ha) and Bentley/Crawford

(1326.67 stems/ha) having similar densities (F = 6.02; df = 3, 110; P < 0.0008) (Table

?). Comparing all woody sapling and shrub percent cover, Big Dune (7.53%) also had the lowest mean percent cover while the remaining three sites Leesville (24.23%),

Mskoda (25.86%), and Bentley/Crawford (35.60%) had similar woody species cover in their plots (F = 7.80; df = 3, 114; P < 0.0001).

Additionally, Big Dune (38.59 cm) contained the largest average diameter trees whereas the other three sites Bentley/Crawford (26.72 cm), Leesville (27.56 cm), and

Mskoda (26.0 cm) had fairly the same average diameter of trees (F = 15.31; df = 3, 104;

P < 0.0001). The size class distributions of all trees in the four sites were very different. 14

The normal bell-shaped curve that represents a single cohort stand was exhibited in Big

Dune (n = 51) where there were no small diameter trees (>25cm) present (Figure ?).

The diameter distribution in Bentley/Crawford (n = 297) is also consistent according to a single cohort stand; however there are substantially more trees compared to Big Dune.

Based on Mskoda’s (n = 170) diameter distribution, results suggest a single cohort stand with 98% of trees between ten to forty centimeters even though a few trees were in the forty to sixty centimeter range (Figure ?).

The combined dendrochronological age structures among the four sites (n = 114) including both the increment cores and cross-sections resulted in 53% of the trees were established before 1957 while the remaining forty-seven percent were less than 50 of age. On average Big Dune (67 years) was the oldest, followed by Bentley/Crawford (57 years), Leesville (52 years), and Mskoda (44 years) (F = 10.18; df = 3, 110; P < 0.0001)

(Table ?). However, there were various recruitment periods across all the sites.

Mskoda’s age structure clearly shows two distinct cohorts around the early to mid

1940’s and mid to late 1960’s (Figure ?). In agreement with the diameter distribution graph, the age structure of Leesville appeared to have multiple recruitment periods ranging from late 1800’s to early 1980’s (figure ?). Bentley/Crawford’s major recruitment periods started in the early 1940’s and lasted into the late 1950’s (figure). Big Dune undoubtedly represented a single cohort stand with 71% of sampled trees established between 1932 to 1945 (figure ?). The age diameter distribution was highly variable across all sites; for example, trees with diameters of 35-40cm ranged in ages from 39 to

128 years (figure ? ).

There were no quantifiable differences in stand health measurements between the sites (figure ?).

15

Fire History

There were only 13 fire free years since 1930; a fire was recorded in a different year, 64 of the 77 years. The cross-sections contained 289 fire scars with the first fire scar recorded in 1907 and the last in 2007 (figure ?). Almost all the trees (92%) were first scared before 45 years of age and 66% of those trees were initially scarred before they were ten years old (figure?). Of the 289 fire scars, 156 (58.6%) of them occurred in the dormant season, 98 (36.8%) in early growing season, 7 (2.6%) in middle growing season, 5 (1.9%) in late growing season, and 23 (8%) fire scars seasonality could not be determined (Figure ?).

From 1930 to 2007, the four sites contained 129 fire intervals that ranged from 1 to 10 years (Table ?). Big Dune had the most fire events (41) with a mean fire interval

(MFI) of 1.68 years and Weibull median fire interval (WFI) of 1.45 years. Mskoda contained the next lowest fire MFI of 2.18 years and WFI of 2.02 years calculated from a total of 33 fire events. Bentley/Crawford had slightly fewer fire events (30) than

Mskoda which was calculated to a 2.47 MFI and 2.24 WFI. Leesville had the least amount of fires (25) with a MFI of 2.68 and WFI of 2.02. Initial observations from the composite fire history graph (figure ?) appears that there may be a relationship between fire frequency and climate since numerous fires occurred during the drier years in the late 1930’s, 1966, 1977, 1988. However, results from superposed epoch analysis (SEA) determined no strong statistical associations (Figure ?).

Kankakee County Cattle Population

The cattle population for Kankakee County has fluctuated since 1926 with the highest head of cattle of 35,000 in the early 1940’s and early to mid 1950’s. Even though there 16 have been fluctuations in the cattle population, the overall population has steadily decreased to 7,300 head of cattle in 2007 (figure ?).

Discussion/Conclusion

• The importance of fire is well known for creating and maintaining oak dominated

systems over the last several thousand years (Abrams 1992).

• The extent of oak savannas within the Midwest was estimated between 11-13

million hectares at the time of European settlement in association with fire.

• Fire frequency of 2 years or less maintains prairie species but does not permit

tree regeneration.

• Fire frequency greater than 2 years permits tree regeneration but canopy closure

occurs in the absence of other disturbance.

• Grazing appears to play a role in maintaining open conditions where fire

frequency is greater than 2 years. 17